Collaborative Shield: Strengthening Access Control with Federated Learning in Cybersecurity
DOI:
https://doi.org/10.63282/3050-9262.IJAIDSML-V5I4P104Keywords:
Federated AI, Cybersecurity, Access Control, Anomaly Detection, Distributed Learning, Zero Trust, Privacy-Preserving Machine LearningAbstract
The rapid evolution of cyber threats has made system security and data protection increasingly critical concerns for organizations worldwide. Federated Artificial Intelligence (AI) offers a promising approach by enabling distributed learning that preserves data privacy while facilitating secure collaboration. This paper explores how Federated AI can enhance access control systems by enabling anomaly detection, policy enforcement, and adaptive threat response in real-time. Traditional centralized AI models require data aggregation at a single location, creating potential breach vectors and compliance challenges. In contrast, Federated AI mitigates these risks by training models across decentralized nodes while maintaining data locality. We present a comprehensive framework implementing robust access control mechanisms that leverage collective intelligence while preserving sensitive information. By integrating Federated AI with Zero Trust principles, we demonstrate a dynamic access control system that adapts to evolving user behaviors and environmental contexts. Our experimental evaluation, using real-world datasets like UNSW-NB15 and CICIDS2017, shows that the proposed framework achieves 93.7% accuracy with strong privacy guarantees (ε=1.0). We discuss key innovations including edge-based real-time anomaly detection, privacy-enhancing techniques such as differential privacy and homomorphic encryption, and the integration of generative models for attack simulation. Finally, we analyze the advantages, limitations, and future directions of Federated AI in cyber defence applications
References
[1] Sharma, V., & Singhal, R. (2022). Comprehensive survey on federated learning architectures for IoT security and privacy. IEEE Internet of Things Journal, 9(14), 12478-12495.
[2] Ahmadi-Assalemi, G., & Al-Khateeb, H. (2007). Dynamic role-based access control models for distributed systems. Electronic Notes in Theoretical Computer Science, 186, 27-42.
[3] Liu, Y., Chen, J., & Wang, D. (2021). Artificial intelligence applications in cybersecurity for connected environments. Discover Internet of Things, 1(1), 12.
[4] Zhang, K., Yang, X., Cheng, H., Li, P., & Wu, Y. (2021). Collaborative federated learning approaches for enhanced intrusion detection in IoT networks. IEEE Access, 9, 142567-142583.
[5] Heinrichs, L., Muller, B., & Peters, T. (2022, November). Implementation frameworks for federated learning in security operations centers. In Proceedings of the 11th European Symposium on Security and Privacy Engineering (pp. 145-152).
[6] Rajesh Kumar, Practical approaches to federated learning for network security, DataScience Journal, online. https://www.datasciencejournal.org/articles/federated-learning-network-security/
[7] Park, S., Lee, J., & Chen, Y. (2018). Access control frameworks for emerging edge computing environments. Computer, 51(10), 86-92.
[8] Advancing Cybersecurity Through Federated Learning: Principles and Applications, SecurityWeek, 2024. online. https://www.securityweek.com/advancing-cybersecurity-federated-learning-principles-applications/
[9] Rodriguez, M., Thompson, K., Vergara, E., & Davis, C. (2022). Federated learning technologies: classification, implementation challenges, and future directions. Electronics, 11(4), 685.
[10] Wang, L., Zhao, S., Liu, B., Yang, Q., & Zhang, Z. (2024). Enhancing network security through attention mechanisms in federated graph neural networks. Scientific Reports, 14(1), 18734.
[11] CyberArk Research, Modern approaches to access control implementation, CyberArk, online. https://www.cyberark.com/resources/access-control-implementation/
[12] Victoria Chen, Transforming security operations with artificial intelligence, Cybersecurity Ventures, 2024. online. https://www.cybersecurityventures.com/ai-security-operations/
[13] Chen, X., Wang, Y., Li, K., Zhang, J., Liu, X., & Wei, J. (2021). Privacy-preserving federated learning for intrusion detection: architectures and challenges. IEEE Transactions on Industrial Informatics, 18(5), 3489-3500.
[14] Michael Evans, Understanding access control systems in distributed environments, CSO Online, online. https://www.csoonline.com/article/distributed-access-control-systems/
[15] Privacy-First AI: Implementing Federated Learning in Cybersecurity, PrivacyTech, online. https://www.privacytech.org/federated-learning-cybersecurity-implementation/
[16] P. Mannem, R. Daruvuri, and K. K. Patibandla, “Leveraging Supervised Learning in Cloud Architectures for Automated Repetitive Tasks.,” International Journal of Innovative Research in Science,Engineering and Technology, vol. 13, no. 10, pp. 18127–18136, Oct. 2024, doi: 10.15680/ijirset.2024.1311004.
[17] Li, H., Chen, D., & Wong, F. (2021). From federated optimization to distributed architecture selection: a comprehensive review. Complex & Intelligent Systems, 7(2), 624-638.
[18] Chen, Y., Xia, H., Roberts, S., Qin, Z., & Fletcher, J. (2022). Design patterns and implementation strategies for federated learning systems. Journal of Systems and Software, 191, 111298.
[19] Park, S., Williams, A., & Taylor, M. (2020, December). Deployment architectures for AI-based security systems: multi-organizational study. In 2020 27th Asia-Pacific Software Engineering Conference (APSEC) (pp. 412-421). IEEE.
[20] Wang, H., Zhang, L., Zhou, Y., Liu, T., & Chang, C. (2022). Balancing utility, privacy, and fairness in distributed learning environments. Computers & Security, 122, 102943.
[21] Patel, N., Johnson, R., Smith, K., Alvarez, M., & Wright, J. (2023). Survey of attack vectors and defense mechanisms in federated learning systems. IEEE Transactions on Pattern Analysis and Machine Intelligence, 46(5), 2656-2671.
